Author: Vincent Diepeveen
Date: 06:35:14 05/22/01
Go up one level in this thread
On May 22, 2001 at 08:40:50, Bertil Eklund wrote: >On May 22, 2001 at 08:10:15, martin fierz wrote: I already posted it. Here a small hastely put together table P3 P4 10+ 20+ minimum number of cycles wasted on branch misprediction 16kb 8kb level 1 code cache (note the p4 has 8kb instruction TRACE cache) 3 3 number of instructions a clock which can be processed theoretical 16kb - level 1 data cache (can't remember the P4 has any) 2 1 number of FPU/MMX decoders, so P3 is 2 times faster here 0.5 3 number of clocks needed for shift instruction 0.5 1 number of clocks needed for multiplication 1Ghz 1.7Ghz the only good thing is that the P4 clocks higher I can't remember any terrain where the P4 was good. Perhaps a future processor can profit from the experience they ahve with new technology. note the p4 is 2 times BIGGEr as the p3 cpu, so it's more expensive to press. I wonder how the K7 performs when it gets 0.13 technology! >>hi, >> >>i just got myself a PIV CPU at 1.4GHz and i'm slightly surprised that my >>checkers program isnt running any faster on it than on a PIII 866 (that's right >>- NO faster, not even 5%). my guess is that the CPU is sitting around waiting >>for data. up to now, from Pentium 150 to to PIII 450 to K7 600 to PIII866 my >>nodes/sec scaled pretty well with MHz, just with this new CPU not anymore. the >>computer has a 100MHz bus architecture, so it's not very fast there, although it >>has fast 400MHz RAM. >>has anyone else observed such a thing with a fast PIV? is it possible that this >>is the bottleneck for my program? is there one component in today's computers >>which is generally the bottleneck? is there a way for me to make an estimate by >>looking at my code if this could be the problem? >> >>cheers >> martin > Hi! > >Yes most chess-programs, performs on a P4-1,4 as an Athlon or P3-800. > >Bertil
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.